'BISP to transition towards dynamic registry and biometric payments,' says Marvi Memon, its Chairperson

Marvi Memon evokes partisan responses on social media. But sitting in her office discussing the intricacies of Pakistan's flagship social-safety net (SSN) programme, BR Research found the former banker-turned-politico adequately apolitical. Arguably, the Benazir Income Support Program (BISP), which Memon has been heading since February 2015, has evolved as a symbol of relative political maturity in the country. Despite a landslide victory in the 2013 General Elections, the PML-N government decided to keep funding the BISP, which was launched by its rival PPP in 2008.

Under the current government, yearly cash grants have more than doubled between FY13 and FY16 to reach Rs 96.65 billion in FY16 (source: BISP Progress Report, 2014-16). Even Memon, who as an opposition member in 2011 had accused the then PPP government of turning poor people into "beggars on BISP forms", has come around to become the program's vocal champion.

BR Research had a detailed discussion with the BISP Chairperson and her technical team on aspects related to BISP's database design and payment disbursement mechanism. Eight years on, after abundant criticism on oft-reported inefficiencies in BISP design and payment process, there is a sense that the organisation has entered a third phase on its path to more precisely identify the poor and effectively disburse welfare payments.

<B>i. Phases 1-2: Take off, identification issues, modification</B>

At the heart of BISP is demographic profiling. When the programme started in 2008 under the PPP-led central government, each MNA/senator and MPA was asked to identify 8,000 and 1,000 poorest recipients, respectively, in their constituencies. That was the first phase of BISP's design. As the program captured donor interest, just two years later, that process - politicised, unscientific, and inadequate as it was - was replaced with a scientific method, with technical support from donors. In phase 2, the Poverty Scorecard Survey (PSS) came out in 2010.

As a result of PSS, which identified eligible households through proxy means test (PMT), only 13 percent of the forms submitted by the politicians could be retained in the new database, Memon mentioned. She said the World Bank had validated the PSS in the years since, declaring it number five among the world's social safety-net (SSN) databases. Based on the PSS, the illustration below summarises the regional break-up of BISP-eligible families. Out of the 7.7 million such households, BISP has a coverage of 5.4 million families as of end 2016.

<B>ii. Phase 3: Reaching for the digital solutions</B>

But the problem with PSS is that it is a static database. It has been more than six years since the scorecard came out. It is possible that some among the hitherto-eligible families may have done better and graduated from the cut-off score of 16.17, which is the PSS ceiling for BISP-eligible households. Moreover, it is possible that the 'new poor' may have been locked out of the database.

<B>a) National Socio-economic Registry</B>

Memon acknowledges such concerns and pointed out towards efforts to develop a dynamic registry. She referred to it as the National Socio-economic Registry (NSER), a static version of which already exists. But in the future, it will be updated on a regular basis and use a better design to weed out issues in identifying the deserving households.

"To be able to identify the poor, we have to profile the whole population first. So, the NSER, unlike the perception about it, is not a poverty census, or a poverty scorecard, or a poverty survey. Under NSER, we are using the PSS to grade the entire population on a scale of 1 to 100, with 1 being the poorest and 100 being the richest families. The database was shared with other government bodies and development institutions, and forty of them have used it to plan their activities," Memon said.

Through further improvements, the idea is to make NSER the number one SSN database in the world. The BISP chief said that on one hand, BISP has engaged technical experts from all provinces and housed them under a dedicated NSER department. On the other hand, politicians have also been kept in the loop to share the lessons learnt in the process. Memon considers local governments as effective force multipliers for the kind of grassroots support a dynamic NSER is going to require.

When asked about the mechanics of phase 3, the BISP chairperson told BR Research that she had put two approaches in motion, to see which approach provided better targeting and more value for money. One is the universal targeting, where household data is collected through door-to-door surveys. The other is the desk approach, where people self-register via Nadra desks that are set up in tehsils and union councils. The desk-approach started in June 2016 in four districts - Haripur, Bahawalpur, Sukkur, and Naseerabad. Memon plans to cover 16 districts by June 2017, and after that, BISP will go for a national rollout, depending on which approach is better.

But the desk approach is not without pitfalls. "So far, we have found that only the poor women showed up because the rich won't waste their time in queues. Also, folks were not coming down from the mountains because they were expecting that like last time, we will go to their doorstep. So this approach risked leaving out a significant portion of population. To counter that, we have used social mobilisation strategy to get people out. And we are holding the exercise over three months to avoid over-crowding on the desks."

When asked about the challenges with the desk approach, Tahir Noor, the DG NSER mentioned that the biggest difficulty was institutional capability, as the desk approach has been done in Pakistan for the first time.

"There were some institutional, human resource issues with NADRA, which is why we were not able to start simultaneously in all union councils as we wanted to. Desks were established gradually. In some UCs, there were desks but none in others, so people from those areas were coming to these areas, not knowing how long the desks would be there, thus risking losing registration. There were also some crowd management issues in the beginning. There was also a social mobilisation issue. There was wrong perception in some areas that if you do this registration, you get BISP cards and benefits. Then, there is a higher chance of folks to underreport their assets, as seen in other countries.

And if you try to do verification simultaneously, there is deterrence in the population," he explained.
On being asked how the "asset under-reporting" issue is being tackled in the field, Noor replied that, "We told people that we will visit their houses to physically verify their information and in case we found under-reporting, we will exclude them from the database for next five years for every social safety-net program, federal or provincial. In any case, BISP will conduct sample-based verification to sort out under-reporting."

How does the NSER chief plan to keep costs in check, given that a dynamic NSER would need to be updated regularly (every two or three years) to account for inclusion error and exclusion error? "We want quality database but we don't want to be too expensive. Last time we spent almost $65 million on the PSS exercise. And this year, the door-to-door approach is expected to cost $85 million. We will know the true cost of desk approach once the results are in from the districts, around May or June. The costs may take us towards a hybrid approach, where we establish desks once door-to-door is complete so that whosoever has not been surveyed can come and give data at the desks. In the future, we can also operate desks at our own offices nation-wide," Noor maintained.

<B>b) Cash disbursements</B>

The single biggest critique on BISP - and this is a criticism that we hear a lot from many respected names in academia and think tanks - is the loopholes in the disbursement process. One end of the problem is that a mafia of sorts has developed that feeds on the financially-illiterate folks' vulnerabilities and that deprives marginalized women from their fair share of government welfare. When asked how she plans to tackle this menace, Memon first recognised that the problem is deep and needs a response.

"In its early years, BISP started by making the disbursements using Pakistan Post. Over time, the Benazir Debit Card (BDC) got traction and it has now become BISP's biggest product, serving a large percentage of 5.4 million households. But I agree, a mafia of middlemen exists that takes those cards from these poor women and earns commission for withdrawing their payments. There are cases where there is a whole chain of middlemen in a single transaction. We officially recognise that. Most BISP recipients are illiterate, they give up their cards voluntarily, and the middlemen get Rs 200/300 as commission. This is unacceptable, and I want the women to get the exact money that the finance division is releasing - Rs 4834 per quarter," she said.

Memon said that BISP was going to eventually move to the biometric verification system (BVS) - composed of CNICs, thumbprints, and Sim cards - using the POS network. The pilot is being done in nine districts, after which it is going to go nation-wide. A main advantage of BVS, she said, was more access points closer to home than ATMs.

<B>c) The BVS regime</B>

Explaining the technical aspects of the new payment system, Sardar Azmat Shafi, BISP's Director General Finance said that the BVS regime would eradicate the problem of agent mafia, which arose from a card-based system. With BVS, impersonation would not be possible, he said. The 100 percent digital transaction would leave a digital footprint in the system - in the Nadra system, on the POS machine, and within the BISP MIS as well. Verification can be done in different ways as there will be reconciliation, confirmation, and transparency on every transaction. With BVS in place, funds can only land in the hands of the woman beneficiary, which was previously not the case.

The DG Finance highlighted a few challenges in the new payment model. "Instead of the debit card instrument, we use the universal ID, which is your thumbprint, making the thumb the withdrawal instrument, thus making the transaction card-less. But, in order to develop it, including making the bank account, there is a lot that needs to be put in place. There are a lot of regulatory requirements. So, we will be able to bring the full system sometime in 2017."

<B>d) A two-way communication</B>

Humaira Zia Mufti, BISP's DG for Cash Transfers noted that another benefit of the new payment model was that BISP could now interact with the beneficiary. "Because there are so many beneficiaries, we needed a digital solution. So, BVS ensures a communication channel. Using the mobile phone, we can now inform the beneficiary of many things. The most important is to let her know that her money has arrived. As of now she doesn't know - she has to go to the AD's office several times, to ask if the money has arrived. Or she goes to the ATM to check. The BVS process will start with a text message, backed by a robocall. Even if she cannot read, she will be able to guess that 'December, March, June, and September are the four times we get money and I heard some message on my phone and this could be that money'."

Further explaining, Humaira said after improving current operations, the long-term initiative was more sophisticated. "On the SIM, the call to helpline is free of cost. Secondly, we are training the beneficiary to use a code. Whenever she enters a code, her account balance will show. She can call the BISP fori raabta or bank helpline and check her balance. For confirmation, they will ask her mother's name, etc, and she can check her balance. SIM is a two-way communication process. Now we can call them and ask questions as well. For almost 600,000 women, data are now coming in via the SIM. Now we will also be able to check how many women have kept the SIM active."

<B>e) Teething troubles</B>

With regards to payments, another problem BR Research pointed out was that some BISP officials who had been accused of embezzlement were merely transferred, and not suspended pending completion of an inquiry. Memon responded that the internal control at BISP - an audit department and monitoring and evaluation teams - had caught such suspects. "We handed them (suspects) over to the FIA. We are not the police," she said, suggesting she did not have the power to suspend unscrupulous BISP officials. Moreover, the fake SMS making rounds was something BISP could only identify and report, she said, but not prosecute on its own.

<B>iii) Predictors of socio-economic status</B>

Noticing BR Research's interest in the PSS variables, Tahir Noor, the DG NSER explained that between 200 and 300 variables were selected from surveys such as PSLM and HIES. "We take those variables, run regressions, and identify those variables which have the largest impact on the socioeconomic conditions of the households. We find that some 20-25 variables fit that profile, so they are included in our PSS. This time we used HIES (2013-14), which is the latest survey available.

When further inquired what has been the most precise proxy so far to gauge poverty status for household, he said it was mishmash. "Education level is one very strong variable; secondly, number of dependants in comparison to number of rooms available in the household (lower the number of rooms, the more the congestion, the higher and the poverty); the type of toilets, or whether there is a toilet."

Six months into the pilot, what is Memon's assessment? "As a politician, my personal assessment is that we will have to work with a hybrid approach, to bring together the desk and door-to-door methods. Both approaches have advantages. The objective is to avoid a situation where we are stuck with a static registry. We want a dynamic registry, where we have data certified every year or every two years. We want to see better-off folks move out, so more poor folks can move into registry without waiting for five years."

About the ongoing fieldwork, the DG NSER said that there were data-entry operators with tablet PCs, who were using specially-designed software to feed the data. The data come to BISP, where it is processed to calculate the poverty scorecard. "So far we have registered 500,000 households in these four districts. After this phase, there will be a targeted performance evaluation survey, to be done with the technical support of the World Bank, to ascertain real poverty in those districts. That will be followed by a door-to-door survey in those four districts. We will compare the results of desk approach and door to door approach with the real poverty survey, on the basis of coverage, target accuracy, and cost-benefit analysis. On the basis of these parameters we will establish value for money for both approaches."

<B>iv) Defining the mandate</B>

In the last sit-down with BR Research, the BISP chairperson also spoke of her desire to expand the conditional cash transfer programs. "It is often overlooked that BISP's Waseela-e-Taleem initiative has helped put in schools 1.3 million children, who were not sent to school before, because of an extra cash transfer. The recipient woman receives an additional Rs 250 per child per month if her child goes to school. Under this conditional cash transfer, Rs 2.87 billion has been disbursed between 2014 and 2016."

When asked about BISP and its questionable impact on female empowerment, Memon noted that BISP now had 50,000 beneficiary committees, which were modelled on the Mexican and Latin American experience. "A beneficiary committee is structured such that there are 25 women per committee, they have one woman leader, and they meet on a monthly basis. I have been telling women empowerment advocates and NGOs that they can use these committees for their advocacy all around the country. But I have noticed that when you, as the government, offer to share such platforms, no one takes you up on that offer," she lamented.

Towards the end, on the question about BISP's exact mandate, Memon explained that she did not see BISP primarily in the field of poverty alleviation, but poverty management via income support.

"Both the finance minister and I have gone on record, as politicians, and have said that BISP is doing poverty management. Government has other graduation models which are doing poverty alleviation. Under the last government, a lot of initiatives were taken under BISP.

We have kept the good parts and I'm open to share our recipients, which I call our customers, with other government programs doing poverty alleviation. For instance, some 50,000 of my customers are enrolled in the Prime Minister's Interest-free Loans Scheme. Its database is with us and we already know where those 50,000 customers are and how they are doing," she concluded.